Materials Science and Engineering B 111 (2004) 49–56

Development of a novel combustion synthesis method

for synthesizing of ceramic oxide powders
Chyi-Ching Hwang∗ , Tsung-Yung Wu, Jun Wan, Jih-Sheng Tsai
Department of Applied Chemistry, Chung Cheng Institute of Technology, Tashi Jen, Taoyuan 335, Taiwan, ROC

Received 12 November 2003; accepted 23 March 2004

Abstract

A novel combustion synthesis method has been developed to prepare electronic ceramic oxide powders—Ni0.5 Zn0.5 Fe2 O4 , ZnO, LiCoO2 ,
BaFe12 O19 and YBa2 Cu3 O7−x (x
0.25). Organic compounds (e.g., glycine, urea, citric acid, alanine, or carbohydrazide) to be mixed directly
with metal nitrates without adding water, is the key technique of this method. Metal nitrates acting as oxidants were also used as cation
sources, whereas an organic compound was employed as fuel. By directly mixing and thoroughly dehydrated, the reactant mixture having
appropriate stoichiometric ratio of metal nitrates to organic fuel would transform into a flammable precursor. Once ignited in the air at room
temperature, the precursor underwent a combustion process and yielded voluminous loose powders. Judging from experimental results, the
feasibility of the proposed combustion synthesis method was proved to be efficient, quick and simple in becoming one of the suitable methods
for mass production of ceramic oxide powder. Furthermore, a relatively low sintering temperature of 950 ◦ C for 2 h is sufficient to sinter the
as-synthesized Ni–Zn ferrite powders up to about 97% of theoretical density due to their size in nano-scale. The as-sintered Ni–Zn ferrite
showed good magnetic properties as characterized by using VSM and impedance analyzer.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Ceramic oxide; Combustion synthesis; Metal nitrate; Organic compound; Magnetic property

1. Introduction variety of wet-chemical methods have also been reported to

Electronic ceramics are widely used as elements and
components in industries such as, electrical engineering,
electronics and information. Not only, have they become
the main stream in the advanced ceramics market, but they
have also been developed well in applications with respect
to both scope and magnitude. Overall, ceramic oxides are
the most important ones among these electronic ceramic
materials [1]. The conventional method to prepare ceramic
oxide powders is to perform the solid-state reaction of ox-
ides and carbonates. These starting materials are mixed at
a specific composition ratio and then calcined in the air or
oxygen atmosphere at a high temperature for a long time.
Intermittent grinding is required to obtain phase-pure and
homogeneous product powders. Although this method us-
ing conventional solid-state reaction is relatively simple,
yet it is time-consuming and energy intensive. Recently, a
∗ Corresponding author. Tel.: +886-33-891716x312;
fax: +886-33-892494.
E-mail address: cchwang1@ccit.edu.tw (C.-C. Hwang).
be effective in generating ultra-fine and more homogeneous
powders of ceramic oxides [2–4]. However, complex sched-
ules and low production rate are the common problems of
these wet-chemical methods.
Fortunately, the drawbacks of these methods as men-
tioned above could be partially eliminated by utilizing the
combustion synthesis method, which has been applied to
the synthesis of various high performance materials includ-
ing ceramics, intermetallics and composites [1,5]. Various
advantages such as low processing cost, energy efficiency
and high production rate [6] are among the benefits of the
combustion synthesis method. With regard to the combus-
tion synthesis of ceramic oxide, the general procedure is
to use corresponding metal powder to be mixed thoroughly
with metal oxides in a proper ratio, and then press them
to get the reactant compact. The compact is ignited in an
oxidizing atmosphere and ceramic oxide products are thus
obtained after combustion [1,7–9]. This indicates that the
metal powder is the fuel whereas the oxidizing agent is
oxygen. Both materials cause combustion and results in
high temperature with phase transformation to obtain the

0921-5107/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.mseb.2004.03.023
50 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56
final product. However, the reaction needs to be conducted
under a required oxygen pressure. At this operating point,
the events of melting of metal, coalescence of melt, and
grain growth of as-synthesized product may occur due to
the combustion reaction arising out of high temperature
(about 1400–1600 ◦ C). Such adverse factors may decrease
the simplicity of procedure, lower the reaction converse,
and reduce the product quality. Therefore, the combustion
synthesis method can be further improved significantly.
One approach that can be used is the examination of gun-
powder reaction. The weight ratio of gunpowder is 75% salt-
peter, 10% sulfur and 15% charcoal. As soon as it is ignited,
a rapid (but over-simplified) reaction takes place [10]:

2KNO3 + S + 3C −
→ K2 S + N2 ↑ +3CO2 ↑
A large quantity of oxygen provided by oxidizing agent
KNO3 is enough to complete the reaction without any exter-
nal provisions. It should be noted that the three ingredients
of gunpowder are solid and, among them, potassium nitrate
acts as an oxidant whereas sulfur and charcoal act as fuels.
With this in mind, we have come up with an idea to de-
velop a new combustion synthesis method for synthesizing
oxide ceramic powders. That is, metal nitrates added with
a suitable fuel powder in an appropriate ratio are ignited
and completely burnt to form ceramic oxides. Characteris-
tically, gunpowder being a special compound can generate
a lot of gases when burned. If such application is utilized
in the synthesis of ceramic powders, a great deal of gases
can be released via the combusting reactants to be expedi-
ently conducive to heat loss through convection, resulting
in rapid reduction of product temperature. On a practical
level, it is anticipated that fine powders could be produced
because of the inhibition from both the growth of grains and
the agglomeration of particles. As stated above, a theory for
improving the combustion synthesis method has been estab-
lished, accordingly.
2. Experimental procedure
Using suitable raw materials in combustion synthesis en-
sures the stability of the chemical composition and high
Table 1
Some properties of organic compounds used in this study
Properties Organic component
Alanine Glycine
Structural formula H2 N–CH2 –COOH
Molecular weight (g/mol) 80.1 90.1
Heat of combustion (kJ/g) 18.2 13.0
Decomposition temperature (◦ C) 314 262
quality of products. More importantly, it must be readily
available and convenient to be used. In addition, it should re-
act non-violently and produce non-toxic gases. This research
selects alanine, glycine, carbohydrazine, urea, and citric acid
as fuels to confirm both the feasibility of our proposal and the
applicability of using these five organic compounds. Some
of their properties from the literature [11,12] are listed in
Table 1. Besides, metal nitrates are utilized in the present
work because of its dual role of being the cation sources and
the oxidant. Analytical-grade organic compounds and metal
nitrates were used as raw materials in this study.
Firstly, this research focused on synthesizing Ni0.5 Zn0.5
Fe2 O4 to be our target product. Ni0.5 Zn0.5 Fe2 O4 , a soft mag-
netic material, can be used in the different devices including
radio electronics and other areas of modern technology due
to their high resistivity and low eddy current [7,8]. The prod-
ucts were synthesized by an amount of 25 g per batch. The
following illustrates the operation procedure. First, weigh
the required quantity of metal nitrates and put it in the alu-
mina crucible. Then, add organic fuel of which it has an
“equivalent stoichiometric ratio” with respect to metal ni-
trates. Afterwards, stir and mix them completely. Here, the
definition of the “equivalent stoichiometric ratio” implies
that the oxygen content of oxidant can be completely re-
acted to oxidize/consume the organic fuel exactly [13]. As a
result, the combustion process can produce CO2 , H2 O and
N2 gases without the necessity of getting oxygen from out-
side. Taking metal nitrates (oxidant) and glycine (fuel) to be
reactants in synthesizing Ni–Zn ferrite as an example, the
following is its stoichiometric chemical reaction:
0.5Ni(NO3 )2 + 0.5Zn(NO3 )2 + 2Fe(NO3 )3
+ 40
9 NH2 CH2 COOH
→ Ni0.5 Zn0.5 Fe2 O4 + 80
9 CO2 + 9 H2 O + 9 N 2
100 56
(∼ 26.2 moles of gases produced)
Metal nitrates possess hygroscopicity, consequently, they
easily absorb moisture and become slurry. Therefore, the
reactants can be mixed homogeneously during the stirring
process. This slurry mixture was heated on hot-plate to dehy-
drate thoroughly. The dried mixture (hereafter to be termed
as precursor) possesses the characteristic of combustion, and
Carbohydrazide Urea Citric acid
75.1 60.1 192.1
12.6 10.5 10.2
153 135 175
 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56
it can be ignited to start combustion reaction by using mini
gas burner. By doing so, flame and smoke are produced,
and the products left behind are voluminous and fluffy in
nature. The flow chart of the process is shown in Fig. 1.
The ferrite powder so prepared as described above was
mixed with an appropriate amount of 5 wt.% polyvinyal al-
cohol (PVA) to act as a binder, and was granulated using a 60
mesh sieve. The granulated powder was uniaxially pressed
at a pressure of ∼180 MPa to form green toroidal specimens
with theoretical density being around 48% and dimensions
being 20 mm in o.d., 10 mm in i.d., and 4 mm in thickness.
The specimens were sintered at 950 ◦ C for 2 h in air with
a heating rate of 10 ◦ C/min and were subsequently furnace
cooled.
Thermogravimetric analysis (TGA; TA Instruments, SDT
2960) of the precursor was carried out at a heating rate of
10 ◦ C/min in static air. Phase formation of product was iden-
tified by using X-ray diffraction (XRD; SIEMENS D5000)
with Cu K␣ radiation. The morphological features of the
product were imaged by transmission electronic microscope
(TEM; Hitachi H-7100) and scanning electronic micro-
scope (SEM; Joel 6400). Powder surface area was measured
using the single point BET nitrogen adsorption method
Fig. 1. Flow chart for synthesis of ceramic oxides using this proposed
combustion method.
51
(Micromeritics, ASAP 2010). Densities of the sintered
specimens were measured using the Archemede’s principle.
The maximum combustion temperature (Tc ) was measured
by Pt–Pt–Rh thermocoule. The carbon and metallic cation
contents in the as-synthesized powders were measured by
CHN elemental analyzer (Perkin-Elmer, 2400(II)) and in-
ductively coupled plasma-atomic emission spectrometer
(ICP-AES; Jarrell-Ash, ICAP 9000), respectively. Vibrating
sample magnetometer (VSM; Toei, VSM-5) was used to
determine the saturation magnetization (Ms ) of the product
at room temperature. Initial permeability (µi ) and quality
factor (Q) of the sintered toroidal specimens were measured
at various frequencies in the range of 10 kHz–13 MHz using
a Hewlett-Packard impedance analyzer model 4191A.
3. Results and discussion
3.1. Thermal analysis of precursor
Fig. 2 shows the typical TGA plots of the various precur-
sors prepared with different organic fuels. In this work, it is
our conception to define an initiation temperature (Tin ) as
a temperature when sample weight starts changing rapidly
during chemical reaction. For example, a significant weight
loss was observed as temperature reached to 125 ◦ C when us-
ing carbohydrazide as organic fuel, indicating that the chem-
ical reaction took place very rapidly (i.e., the slope of weight
loss–temperature curve is very steep). At around 130 ◦ C, the
remaining weight is only 22% of its original weight, and the
slope begins to level off gradually. While using glycine and
alanine as organic fuels, the thermogravimetric analyses of
both precursors demonstrate much the same results. Both
precursors have Tin at around 180 ◦ C, and their reactions are
very fast. Both precursors stop their changes of weight loss
at around 180 ◦ C. Their weight loss is close to 83%. When
using urea as an organic fuel, its Tin is around 165 ◦ C, but
the change of weight loss is much flatter as compared with
Fig. 2. Typical TGA plots of the precursors prepared by using various
fuels.
52 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56

glycine and alanine. As temperature reached to 280 ◦ C, its

weight was reduced to 18% of its original weight. It is shown
that the combustion phenomenon pertaining to precursors
of using aforementioned three organic fuels are much more
drastically than that of using urea as organic fuel. This re-
sult is in consistency with the combustion phenomenon of
using urea as fuel—the combustion reaction of the precur-
sor is not rigorous. It can be inferred that urea with rela-
tively low combustion heat, is the cause that contributes to
this outcome. When using citric acid, the thermogravimetric
analysis of this precursor indicates that it is quite peculiar.
As the precursor was heated, thermal decomposition started.
From the curve of thermogravimetric analysis, it is observed
that there appears a phenomenon of three-stage weight loss.
Obviously, it belongs to a multi-stage reaction. The weight
loss was not stopped until the temperature reached to 500 ◦ C
(weight loss:83%). When comparing with other organic fu-
els the slope of this curve is relatively flat, signifying that
the reaction is more moderate. This overcome is consistent
with the results obtained using thermogravimetric analysis,
which can be attributed to the reasons that the four organic
fuels that were used beforehand contain amine group that
can be decomposed to produce ammonia under low temper-
ature. In actual fact, reaction between ammonia and NOx
from the decomposed metal nitrates took place and it is writ-
ten in the following [14]: Fig. 3. XRD patterns of the as-synthesized Ni–Zn ferrites prepared by
using various organic fuels: (a) alanine, (b) glycine, (c) carbohydrazide,
4xNH3(g) + 10NOx(g) (d) urea, and (e) citric acid.

→ (5 + 2x) N2(g) + 6xH2 O(g) + 2xO2(g)

The energy released from the oxidation–reduction reaction
causes the combustion reaction to accelerate.
3.2. Phase analysis of product
Fig. 3 shows the XRD patterns of as-burnt products syn-
thesized by using various types of organic fuels. As in-
dicated by the characteristic peaks of Ni–Zn ferrite, the
Ni–Zn ferrite could be produced using each of the organic
fuel studied here. When using glycine, alanine and carbo-
hydrazide as organic fuels, the product seemed to contain
single-phased Ni–Zn ferrite because only the characteristic
peaks of Ni–Zn ferrite were observed. While using urea,
citric acid as organic fuels, the XRD patterns indicate that
there exist other weak diffraction peaks caused by impuri-
ties in the as-synthesized product. Additionally, there exists
a considerable broadening phenomenon in Fig. 3 in terms
of characteristic peaks. This occurrence may be attributed
to the fact that the as-synthesized powder is relatively fine.
Fig. 4 shows the XRD patterns of five as-sintered samples
corresponding to various organic fuels. Notably, sharp and
intensive peaks of a pure Ni–Zn ferrite phase can be seen
from all of the XRD patterns.
3.3. Characterization of materials
Table 2 lists the analytical results obtained from synthe-
sizing Ni–Zn ferrite products using various organic fuels.
According to the calculation using Scherer formula [15], all

Table 2
Effects of various organic compounds and particular properties of Ni–Zn ferrite synthesized by the proposed combustion synthesis method
Fuel Tm a (◦ C) Amount of gas Crystallite Surface area Carbon Ni2+ :Zn2+ :Fe3+ Ms c
produced (mole) sizeb (nm) (m2 /g) content (wt.%) (molar ratio) (Am2 /kg)
Alanine 1245 20.7 38.6 24.7 1.64 0.500:0.467:1.920 60.8
Glycine 1150 26.2 32.7 31.2 1.53 0.500:0.471:1.922 62.4
Carbohydrazide 1380 24.0 43.7 20.6 1.87 0.500:0.462:1.917 58.5
Urea 785 30.7 20.2 48.5 3.82 0.500:0.483:1.936 57.2
Citric acid 725 26.2 22.7 44.1 5.75 0.500:0.490:1.947 55.8
a Tm : the maximum combustion temperature, measured by Pt–Pt–Rh thermocouple.
b Crystallite size of the as-synthesized Ni–Zn ferrite powders calculated from the line broadening of the (3 1 1) XRD peak by Scherrer formula.
c Ms : the saturation magnetization of the sintered Ni–Zn ferrite samples (950 ◦ C/2 h).
 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56 53

Fig. 4. XRD patterns of the sintered samples from different Ni–Zn ferrite
powders synthesized by using various organic fuels: (a) alanine, (b)
glycine, (c) carbohydrazide, (d) urea, and (e) citric acid.

products obtained in this work are nanocrystalline with their

sizes ranging between 20.2 and 43.7 nm. The surface areas
of the synthesized products range from 32.1 to 48.5 m2 /g.
The carbon content of the products synthesized using urea
and citric acid is higher than that using other fuels, since the
heat released during combustion is relatively lower for citric
acid and urea (see Table 1). It was found that the maximum
temperatures during combustion reaction were relatively low
for both these two precursors, which might have caused in-
complete reaction. From the result of the analysis using an
ICP spectrometer, the measured Ni to Zn to Fe cation ra-
tio (Ni2+ :Zn2+ :Fe3+ ) in the as-synthesized Ni0.5 Zn0.5 Fe2 O4
was nearly equal to the anticipated ratio, indicating that the
products possess good chemical homogeneity. The satura-
tion magnetization (Ms ) values of the sintered Ni–Zn ferrites
are in the range of 55.8–62.4 A m2 /kg, which are compara-
ble with those obtained using other methods [8,9,16,17]. Fig. 5. Typical SEM photographs of the as-synthesized Ni–Zn ferrite
When using glycine as organic fuel, Ni–Zn ferrite ob- product: (a) thin flakes, (b) clusters of tiny particles, and (c) whiskers.
tained through combustion synthesis possesses better qual-
ity. Also, the price of glycine is inexpensive. Therefore, pestle. There are three major morphologies: (1) one of them
glycine is selected hereafter to acquire the ferrite. shows thin flakes with holes (Fig. 5(a)); (2) the other is the
clusters of tiny particles with loose appearance (<0.1 ␮m),
3.4. Microstructure and sinterability as shown in Fig. 5(b). Both morphologies can be attributed to
the liberation of a large amount of gases during combustion
Fig. 5 shows typical SEM photographs of the as-synthe- reaction; (3) the third is whisker, as shown in Fig. 5(c). Since
sized products after grinding for 3 min by using mortar and there is a formation of whisker, it is our predication that
54 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56
Fig. 6. TEM photograph of the synthesized Ni–Zn ferrite powders.
the combustion synthesis reaction would contain mechanism
of gas-solid reaction. Fig. 6 is a TEM photograph taken
after the as-synthesized product had been fully grinded and
treated with oscillation procedures. The sizes of the spherical
particles are dispersed with negligible agglomeration and
are homogeneously distributed ranging from 30 to 40 nm,
which is about the same as the sizes estimated by using XRD
method.
A quite dense micro-structure can be seen from Fig. 7,
showing the SEM photograph for the fractured surface of
ferrite sintered at 950 ◦ C for 2 h. The texture of the sintered
sample is homogeneous consisting of polyhedral grains with
Fig. 7. SEM photograph taken on the fractured surface of sintered Ni–Zn
ferrite sample.
Fig. 8. Magnetization vs. magnetic field plot for Ni0.5 Zn0.5 Fe2 O4 synthe-
sized by this combustion synthesis method.
sizes 0.5–2 ␮m. The sintered density was found to be 97% of
its theoretical density, indicating that the synthesized Ni–Zn
ferrite can be densified at 950 ◦ C, which is low as compared
to conventional ceramic method (>1120 ◦ C [18]). This indi-
cates that the nano-sized powder obtained by this proposed
method possesses a relatively high sintering activity.
3.5. Magnetic properties
The magnetization (M) of Ni0.5 Zn0.5 Fe2 O4 as a function
of magnetic field (H) is shown in Fig. 8. The magnetization
curve of the as-synthesized ferrite at room temperature was
characterized by lower magnetization value and a single-line
M–H curve, indicating its superparamagnetic nature. Notice
that the as-formed ferrite did not attain saturation value
even at 15 kOe. However, on sintering the ferrite at 950 ◦ C
for 2 h, saturation magnetization was attained. This effect is
obviously caused by the dependence of particle size on Ms .
Apparently, the magnetic saturation moment reaches a max-
imum value of 62.4 A m2 /kg for sintered Ni0.5 Zn0.5 Fe2 O4 .
This can be explained on the basis of spin arrangement
out of Yafet–Kittel type [19]. Fig. 9 gives the typical ini-
tial permeability and quality factor versus frequency of the
Ni–Zn ferrite sample, which had been sintered at 950 ◦ C for
2 h. Flat permeability profile with the value of ∼125 is ob-
served at various frequencies up to 10 MHz for this sample,
exhibiting good frequency stability. Then, the permeability
increases with increasing frequency (10–13 MHz), which
indicates the onset of resonance [20]. It is also observed
that the curve of quality factor (Q) of the sintered sample
exceeds 55 in a broad profile in a frequency range around
2–5.5 MHz.
3.6. Synthesis of other ceramic oxides
Fig. 10 shows XRD patterns, utilizing our proposed
method to synthesize different types of ceramic oxides, by
 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56
Fig. 9. Initial permeability and quality factor vs. frequency for the sintered
Ni–Zn ferrite sample prepared by this combustion synthesis method.
Fig. 10. XRD patterns of other ceramic oxides prepared by using
glicine as fuel: (a) as-synthesized ZnO, (b) as-synthesized LiCoO2 ,
(c) as-synthesized SrFe12 O19 , (d) as-synthesized SrFe12 O19 calcined
at 1200 ◦ C/2 h (inset), (e) as-synthesized YBa2 Cu3 O7−x and (f)
as-synthesized YBa2 Cu3 O7−x calcined at 900 ◦ C/12 h (inset).
55
using glycine as the fuel. As can be seen, pure ZnO (a type
of semiconductor ceramics) and LiCoO2 (to be used as
the electrode of fuel cell) can be directly synthesized from
the proposed method. However, when using this proposed
method to synthesize SrFe12 O19 (a type of hard magnetic
material) and YBa2 Cu3 O7−x (a type of superconductor),
in addition to the target product produced, there exist
other by-products in the as-synthesized product. These two
as-synthesized powders were calcined at 1200 ◦ C for 2 h and
900 ◦ C for 12 h, respectively. In this way, pure SrFe12 O19
and YBa2 Cu3 O7−x could be obtained (XRD patterns of
the two calcined products are also inset in Fig. 10). Even
so, this method is relatively simple, quick, efficient and
economic when compared to solid-state reaction method
and other wet-chemical routes. Evidently, it can be widely
employed to synthesize different ceramic oxide powders.
Other associated studies such as the effects of both molar
ratio of fuel-oxidant and combustion temperature, the inves-
tigation of reaction mechanism, the characteristic analysis
of products etc. are being further investigated.
4. Conclusion
In this work, we have utilized an indicative and precur-
sory experiment strategy to prove the feasibility of our pro-
posed method. Reactants of metal nitrates and organic fuels
pursuant to our proposed combustion synthesis method are
utilized to produce ceramic oxide powder. Although cer-
tain initial products need further calcination procedure to
attain pure phase product, this proposed method possesses
special features such as time-saving, easy to operate, and
energy-saving when comparing with other methods. It is
expected that the advantages of this method would impel
it to become one of the suitable methods for mass produc-
tion of ceramic oxide powder. Moreover, the as-synthesized
Ni–Zn ferrite powder with nano-sized crystallites has a high
activity of sintering and it is capable of being densified at
950 ◦ C for 2 h. The sintered Ni–Zn ferrite produced from
the as-synthesized powder has small-grained microstructure
feature and possesses good frequency stability, high-quality
factor (>55 in 2–5.5 MHz) as well as low permeability value
(∼125 below 10 MHz).
Acknowledgements
Support for this research by the National Science
Council of the Republic of China under Grant No. NSC
92-2214-E-014-002 is gratefully acknowledged.
References
[1] P.B. Avakyan, M.D. Nersesyan, A.G. Merzhanov, Ame. Ceram. Soc.
Bull. 75 (1996) 50.
56 C.-C. Hwang et al. / Materials Science and Engineering B 111 (2004) 49–56
[2] V.V. Pankov, M. Pernet, P. Germi, P. Mollard, J. Magn. Magn. Mater.
120 (1993) 69.
[3] P.S.A. Kumar, J.J. Shrotri, C.E. Deshpande, J. Appl. Phys. 81 (1997)
4788.
[4] A. Dias, R.L. Moreira, N.D.S. Mohallen, J. Magn. Magn. Mater. 172
(1997) L9.
[5] Z.A. Munir, Ceram. Bull. 67 (1988) 342.
[6] J.F. Crider, Ceram. Eng. Sci. Proc. 3 (1982) 519.
[7] Y. Choi, N.-I. Cho, H.C. Kim, Y.D. Hahn, J. Mater. Sci. 11 (2000)
25.
[8] Y. Li, J. Zhao, J. Han, Mater. Res. Bull. 37 (2002) 583.
[9] C.H. Peng, C.C. Hwang, C.K. Hong, S.Y. Cheng, Mater. Sci. Eng.
B 107 (2004) 295.
[10] M.S. Russell, The Chemistry of Fireworks, Royal Society of Chem-
istry, UK, 2000, p. 13.
[11] R.H. Perry, C.H. Chilton, Chemical Engineering Handbook, seventh
ed., McGraw-Hill, New York, 1997.
[12] J.A. Dean (Ed.), Lange’s Handbook of Chemistry, 15th ed.,
McGraw-Hill, New York, 1998.
[13] S.R. Jain, K.C. Adiga, V.R. Pai Verneker, Combust. Flame 40 (1981)
71.
[14] F. Li, K. Hu, J. Li, D. Zhang, G. Chen, J. Nucl. Mater. 300 (2002)
82.
[15] H.P. Klug, L.E. Alexander, X-ray Diffraction Procedures for Poly-
crystalline and Amorphous Materials, Wiley, New York, 1997,
p. 637.
[16] Y. Hayashi, T. Kimura, T. Yamaguchi, J. Mater. Sci. 21 (1986)
2876.
[17] P.S. Anil Kumar, J. Appl. Phys. 81 (1997) 4788.
[18] K. Majima, M. Hasegawa, S. Katsuyama, S. Mishima, H. Nagai,
Mater. Trans., JIM 36 (1995) 1392.
[19] N.S. Satyamurthy, M.G. Natera, S.I. Youssef, R.J. Begum, C.M.
Srivastava, Phys. Rev. 181 (1969) 969.
[20] J.L. Snoek, Physica 14 (1948) 207.